The Perils of Parallels (Lecture)
James R. Davila
This lecture presents my own working ideas on the concept of parallels and their proper use. It is very much work-in-progress and is not intended to be take as a complete or definitive treatment of the topic. I welcome all comments and criticisms and hope that it will generate some discussion on the list. The topic is particularly apt for this course at this time, since several of the upcoming essay topics deal with parallels between the Dead Sea Scrolls and something else and, of course, the international conference on the Scrolls to be held in St. Andrews this June will focus on parallels between the Scrolls and both other forms of postbiblical Judaism and early Christianity
Although parallels are constantly being adduced and held up as proof for one thing or another in biblical studies and related disciplines, I have been able to find very little theoretical literature on the subject. Nearly forty years ago, Samuel Sandmel published his SBL presidential address for 1961 under the title "Parallelomania," which he defined as "that extravagance among scholars which first overdoes the supposed similarity in passages and then proceeds to describe source and derivation as if implying literary connection flowing in an inevitable or predetermined direction" (p. 1). His article remains very useful but I think the discussion can be carried further today. More recently, Jonathan Z. Smith has published a book that explores the biases passed down in critical biblical scholarship by its Protestant (and radically anti-Catholic) origins and heritage, in the course of which he devotes a good deal of discussion to the subject of good and bad parallels. I am unaware of other theoretical discussions of the subject and would be very grateful to anyone who can draw any to my attention.
In this lecture I focus specifically on parallels found in texts, but with some attention to their implications for the social context and background of these texts. Parallels in material culture are largely outside the scope of my discussion, although I will refer to the important article by Frank Moore Cross on the typology of scripts and pottery. My examples focus on biblical studies and Qumran studies, areas which are filled with not only conceptual but also theological minefields.
TYPES OF PARALLELS
Here I set out briefly four types of parallels (with some subtypes) which are commonly adduced in biblical and Qumran studies.
1. Linguistic Parallels
Lexicography (the process of defining words in lexicons or dictionaries) faces many difficulties when applied to ancient languages and must often rely on parallels in attempting to define words. This is especially true when one is trying to define a word that only appears once in a given language corpus (a _hapax legomenon_) or one is trying to decipher a language on the basis of a limited number of archaeologically recovered texts. Lexicographers draw parallels within a language between different forms--say, a noun and a verb--which use the same root; between the uses of a word that appears rarely in early texts but more frequently in later ones; and between parallel forms in cognate languages.
The study of _hapax legomena_ gives many examples of such lexicographic comparisons. _Hapax legomena_ in the Hebrew Bible are frequently compared to uses of the same root in later Hebrew or in Aramaic or even in Arabic (see Greenspahn for many examples). Lexicographic parallels are especially important in the study of Ugaritic, the dialect of a corpus of Northwest Semitic texts inscribed in alphabetic cuneiform on clay tablets, excavated at the site of Ras Shamra in Syria and dating to the second half of the second millennium BCE. Ugaritic is a dialect closely related to biblical Hebrew and often the first insight into the meaning of a word comes from its cognate in Hebrew. Arabic cognates are frequently also applied, although the linguistic distance between Ugartic and Arabic makes these methodologically more problematic.
Linguistic parallels are frequently sought between unrelated languages. For example, Hebrew, Aramaic, and Greek were all spoken by Palestinian Jews at the turn of the era and connections between these groups are often sought across the language barrier. For example, Brian Capper has argued that a difficult Greek phrase in Acts 2:47, _PROSETI/QEI TOU\S SW|ZOME/NOUS KAQ (HME/RAN )EPI/ TO'\)AUTO/_, literally, "he (the Lord) was adding daily together those being saved," has a Hebrew Vorlage behind it. Retroversion on the basis of the LXX supports the possibility that the Hebrew original read "he was adding daily to the Yahad those being saved." In other words, the technical term used by the Qumran sectarians to mean "community" may have been used by the early Jerusalem church with exactly the same sense. The same term may lie behind the Greek of Acts 2:44.
This sort of parallel is important for one aspect of study of the Gospels: the attempt to isolate Greek passages in the Four Gospels which show signs of being translated from a Semitic language, presumably Aramaic. Such attempts are worth making, although they raise formidable methodological problems in that they can only be successful to the degree that the translator produced a less-than-idiomatic translation.
2. Verbal Parallels
By "verbal" I mean parallels that show such close identity of wording (in the same language or in translation) that one can reasonably conclude that there is a direct relationship between the texts.
Quotations are the clearest indicator that one work knew another. As an example, I note the passage in _Apocalypse of Peter_ 4:7-9, which Richard Bauckham has pointed out bears a striking similarity to a passage in _4QSecondEzekiel_. This passage paraphrases Ezek 37:7-8 and the quotation in the _Apocalypse of Peter_ follows the wording of the Qumran text, making it quite likely that the former work knew and quoted from the latter.
2.2 Use of Technical Terminology
In the case of shared technical terminology, the verbal parallels are not quotations but rather words used in a distinctive, technical way, such that one can reasonably conclude that the works have some close relationship that remains to be explored and defined. In earlier lectures for this course I have drawn attention to technical terms that appear in some Qumran texts and that seem to indicate a shared sectarian consciousness to a greater or lesser degree (e.g., "Yahad," "sons of light/darkness," "Belial," "pesher," "prince of light," "sons of truth," "Kittim," "seekers of smooth things," etc.).
New Testament scholars have also noted a long list of terms shared by the New Testament and the Qumran literature. These include "works of the law," "sons of light," Belial," "the lot of the holy ones," and many others. The significance of these is a difficult question that must be taken up separately with each relevant work in the New Testament.
3. Conceptual Parallels
These are connections between texts that fall short of verbal identity but which still may indicate a close relationship.
3.1 Parallels in Structure and Theme
Sometimes two texts can share so many motifs and structural elements that, even though they show no verbal parallels, one is led to suspect that one knew and drew upon the other. Such cases can become very complex. I note a single example. Both David Suter and I have discussed the relationship between the _Similitudes_ of Enoch (_1 Enoch_ 37-71), a Jewish work of the Hellenistic period (probably the first century CE, more or less), now preserved only in Ethiopic, and _3 Enoch_, a work preserved in multiple recensions in its original Hebrew. The central core of the latter, _3 Enoch_ 1-15, seems to have been in existence by the ninth century CE at the latest. As the titles indicate, both works have much to do with the biblical patriarch Enoch and his ascent to heaven and divinization there. There are no verbal parallels between the two works but they share structural elements and strikingly similar motifs, and it is likely that there is a genetic relationship between them. The nature of this relationship is not yet clear. It is possible that the writer of the core of _3 Enoch_ had access to the (Hebrew?) original of the _Similitudes_, but it is also possible that the writer only knew a version of the story via oral tradition, or the truth may lie somewhere between these two extremes.
More attenuated parallels of structure and motif can indicate a shared "theme," rather than direct dependence of one work upon another. Numerous "mythic" and "epic" themes in the Hebrew Bible appear many centuries earlier in the Ugaritic texts. Such themes are a conglomeration of story elements that tend to appear together again and again, even though the contexts and characters change. A single example must suffice. The epic theme of the birth of a patriarch in the biblical tradition includes a number of elements, including: a prospective father who offers a sacrifice (and sometimes an incubation ritual) asking God for children; the promise of children from God in a vision or dream; a barren wife who must contend with a rival for her husband's affection; or a bride who must be acquired by the prospective father on an epic quest. A mixture of these elements precedes the birth of the hero in the stories of Isaac, Jacob, Samson, and Samuel. Moreover, versions of the same theme are recognizable many centuries earlier in the Ugaritic epics of Kirta and Aqhat and many centuries later in the story of the birth of John the Baptist in the Gospel of Luke. (See the works by Cross and Hendel for more on epic and mythic themes in ancient Israelite literature.) The scholarly tool of "Form Criticism" also deals more generally with parallels of structure and theme.
3.2 Parallels in Ideas
One can find parallels in ideas between two works without them having verbal or structural parallels. An example is the frequently noted use of dualistic light and darkness imagery in both the Dead Sea Scrolls and the Gospel of John. More recently proposed cases include the proposed parallels between Qumran sectarian halakhah and the halakhah of the Sadducees (by Lawrence Schiffman and Joseph Baumgarten--see the relevant section of the annotated bibliography for this course), and proposed parallels between Qumran texts and the later Hekhalot literature which I and others have defended.
These three examples illustrate three possible frameworks into which parallels in ideas may fit (I do not doubt that there are others). They may be single ideas shared by two works or corpora which are used to argue a social or genetic connection of some sort. They may be a cluster of ideas in one corpus which correspond to a similar or identical cluster in another corpus of the same period and which is used to argue for the identity of the groups that produced the cluster. And they may be a cluster of ideas in one corpus which correspond to a similar or identical cluster in a much earlier or much later corpus and which is used to argue for a genetic connection of some sort between them.
Models are a tool used primarily in social anthropology to set up cross-cultural categories of human experience and institutions for use in studying specific cultures. An example of such a use of a model is Philip Esler's comparison of the dualisms found in the Qumran Community Rule and in the Gospel of John in light of sectarian theory as formulated by Bryan Wilson. Esler concludes that parallels between the dualisms of John and the Community Rule should be explained not by a genetic connection but by independent adoption of an "introversionist" sectarianism that arose naturally from each group's conscious and deliberate separation from the outside world.
SOME CRITERIA FOR THE EVALUATION OF PARALLELS
What makes a parallel good or bad, valid or invalid? I propose here a number guidelines and principles which I myself find helpful in evaluating and formulating parallels.
1. Make clear what is being compared to what and how.
Here I follow Jonathan Z. Smith who, drawing on "resemblance theory" points out that the phrase "_x_ resembles _y_" does not make sense on its own but can only be shorthand for a longer statement such as "_x_ resembles _y_ more than _z_ with respect to ...," which makes clear what is being compared and to what it is being contrasted (p. 51). Beware of arguments that claim to prove that something is "unique" (e.g., the resurrection of Jesus; the concept of salvation in Christianity; monotheism in ancient Israel). Such claims are ultimately vacuous--everything is unique or it would be the same as something else--and almost always have a theological agenda behind them (e.g., monotheism in ancient Israel is unique among ancient Near Eastern religions and therefore must be divinely revealed).
2. A "parallel" is not necessarily a "borrowing" and, if it is, the direction of borrowing must be demonstrated, not assumed.
Sandmel discusses this point in his article and I largely follow him here. In some cases, such as the shared terminology between the Qumran sectarian literature and the New Testament, we may be dealing with shared ideas in a common culture. In others, one may argue for literary dependence between two works without being sure which used which. In still others, one can argue that one work is quoting another or (more problematic) using another. As an example of the last possibility, I and others have raised the possibility that the Parable of the Sheep and the Goats in Matt. 25:31-46 knew the _Similitudes_ of Enoch and drew upon it. If there is a literary relationship between the two works, it almost certainly involves Matthew's use of the _Similitudes_ and not the other way around. Either is possible chronologically, but it is easy to imagine a follower of Jesus drawing on the Jewish _Similitudes_ for themes of divine mediation to apply to Jesus and much more difficult to imagine a Jewish writer of the late first-century CE expanding one of Jesus' parables for inspiration on the same subject.
3. Take into account both similarities and differences.
This principle follows naturally from the first. If we are clear on what is being compared to what and how, we naturally will want to know where and how the resemblance breaks down. The ways that two similar things do _not_ resemble one another are often very illuminating. The critical tool of "redaction criticism" is designed to extract important information from such failures to resemble. For example, the places where the parallel passages in the Synoptic Gospels differ are likely to show us some of the ways a particular Gospel writer adapted and altered the sources for the writer's own purposes and therefore to give us insight into what those purposes were. In general, a parallel only has meaning when placed in an overall context of differences.
4. The compared elements must be understood in their own cultural and linguistic contexts.
This is a specific aspect of principle three which is especially important for biblical and Qumran studies. An element, or even a pattern of elements, in a text may seem at first glance to be the same as an element or pattern in some other text from another place and time. Yet closer analysis may show that the function of the element or pattern is entirely different in the two texts or the two cultures, in which case the validity and usefulness of the parallel must be called into question.
A famous example is the alleged parallel of the institution of "wife-sistership" in the stories in the Book of Genesis and in Akkadian legal texts from Late Bronze Age (Hurrian) Nuzi. In Genesis 12, 20, and 26 there are episodes (a nice example, incidentally, of an epic theme) in which a biblical patriarch and his wife visit foreigners and the patriarch falsely claims that his wife is his sister. E. A. Speiser noted that in some of the Nuzi texts a man both marries a woman and adopts her as his sister, evidently in order to strengthen the marriage bond by combining marital and fraternal powers and privileges for both spouses. Speiser argued that the stories in Genesis preserved a distant and garbled memory of this institution of wife-sistership (pp. 91-94) but it is now generally recognized that this theory is very difficult to maintain. In Genesis, the ungallant lie about the sistership of the wife is explained quite adequately by the husband's fear that the barbarians around them will do away with him in order to get at his beautiful wife. There is no hint of an idea that any legal fiction is involved and the Hurrian custom of wife-sistership does nothing to illuminate the Genesis accounts. The parallel rests on a chance similarity that has very different meanings in the two contexts.
5. Patterns of parallels are more important than individual parallels.
A single parallel between two texts may or may not indicate an important connection between them. As a methodological principle, it seems to me that it is good to assume that one or two isolated parallels are insignificant and coincidental unless there are strong reasons to believe otherwise. Such a principle would have saved a generation of scholars from the dead end of patriarchal wife-sistership and other similar comparisons between ancient Near Eastern literature and the patriarchal stories in Genesis.
Another example is the comparison of the dualistic use of the motif of light and darkness in Qumran sectarian literature and the Gospel of John, mentioned above. Richard Bauckham has shown that this single point of comparison means very little, inasmuch as the two dualisms are otherwise quite different from one another, the distinction between light and darkness is the most obvious dualism of physical reality, and the Hebrew Bible and other Jewish texts also make metaphorical use of this obvious distinction. Certainly this isolated parallel falls far short of demonstrating a genetic connection between the Qumran sectarians and the Johannine community or even that one group knew the other. The larger and more complex the pattern of parallels, the more we should take them seriously.
6. The more widely shared the parallel, the more general (or vacuous) its significance probably is.
Going back to Smith's formulation, "_x_ resembles _y_ more than _z_ with respect to ...," we see that a similarity between _x_ and _y_ is only significant insofar as it also corresponds to a difference between _y_ and _z_. Therefore, if _x_ resembles _y_ but also _z_ and _q_ with respect to the same thing, one begins to wonder if the similarity to _y_ can be very important. An example of this sort of problem has arisen already in this course regarding the points of comparison between the Essenes and the Qumran sectarian literature. Taken alone, they seem to make a very convincing case for the identification of the two groups. But Golb has pointed out that many of the same parallels can be found in the rabbinic Havurot and Klinghardt has noted a different but overlapping set of parallels with Hellenistic voluntary associations. These new points of comparison do not necessarily invalidate the traditional comparison with the Essenes but they do require us to look closely at each parallel to see how limited it is to the Essenes and also to consider whether it might be useful to construct more general cultural and cross-cultural models for comparison.
In general, if a parallel or pattern of parallels between two texts or corpora or social groups turns out to be a shared element or pattern with much the same function in many texts, corpora, or social groups, one should be cautious of making much of the original comparison between just the two. All the parallel elements need to be evaluated together and it may be that the most valid description will involve them all or even that we are dealing with a trivial phenomenological parallel that in the end does not advance our understanding.
7. Beware of comparisons that imply an evolutionary goal.
The idea that history flows toward an evolutionary goal is one of Hegel's less useful and more enduring contributions. The fact is that we cannot predict the future except on a statistical and very limited basis, and even then never with certainty. Karl Popper has shown that the idea of a utopian system that guides society inexorably toward an inevitable goal is not just extremely problematic, it is incoherent. There is no ³psychohistory² (with apologies to Isaac Asimov).
In biblical studies, the idea of an evolutionary goal is often smuggled in--almost always implicitly rather than explicitly--on the basis of theological assumptions, usually along the lines that pagan religion was guided providentially toward biblical and therefore true religion; that imperfect Judaism was guided providentially in preparation for the revelation of perfect Christianity; and that pristine Christianity was corrupted into Catholicism but providentially restored as Protestantism. These three covert evolutionary lines underlie a great many discussions of parallels and inevitably distort the discussions.
Biblical scholarship has made a fair amount of progress is rooting out the assumption of such evolutionary systems but they are far from gone. A look at a range of textbooks currently used in many university classrooms will quickly verify how much remains to be done. Consider how many textbooks still distinguish corrupt ³Canaanite religion² from its purified descendant ³Israelite religion² and how many still comfortably refer to the ³Intertestamental period.²
Smith's book takes up the third aspect of the evolutionary schema, the corruption of pristine Christianity into Catholicism, and shows how this Protestant ideology has created a great deal of confusion in the study of early Christianity and the religions of late antiquity.
Here I must distinguish such evolutionary schemas from the typological method, a method of ordering data which is quite useful in the sciences and in history. Typology involves ordering specific, generally material, data in its order of development. The styles of ancient pots or ancient alphabets changed gradually, step-by-step, over time and attention to the details of change can give us valuable information about how and when the pottery and letter forms developed. People frequently speak of the "evolution" of pottery styles and alphabets but this is an imprecise way of putting it. Typology differs in important ways from evolution in the sense I have been using it here. First, typology concentrates on very specific and concrete changes rather than trying to establish an evolutionary framework for a whole culture. Second, typology studies gradual changes over time but there is no sense that these changes are tending toward a goal. For more on the typological method, see the article "Alphabets and Pots" by Cross.
8. Beware of unfalsifiable parallels.
As with all scientific--and, I would maintain, historical--theories, it is important to try to formulate our parallels in such a way that they are subject to falsification, that is, that new evidence or a better understanding can in principle show them to be wrong. I grant that in history this is not always easy, or even possible, to do. Some would say it is not even desirable but I hold that it is the ideal up to which we should hold our theories. It seems to me that falsifiability is often a reasonable standard in the quest for parallels. A definition of a word, an attribution of a quotation, or even a pattern of literary influence are usually concrete enough that one can evaluate them against the evidence and imagine getting a different answer with new evidence or better methods. I think it is fair to say that the more abstract the conclusion being drawn from the parallel, the more difficult is the falsification. Cross-cultural models are especially difficult, since they must be general enough to apply cross-culturally but specific enough that it means something when an example in a particular culture corresponds to the model. I would cite the use of Freudian psychoanalysis and Jungian analytic psychology as especially problematic, since both systems have an enormous and perhaps inexhaustible ability to absorb data and explain the data in terms of the system. This is not, as might appear at first glance, a desirable trait, since a theory that can explain all conceivable data cannot be shown to be wrong.
WHAT CAN PARALLELS PROVE OR ATTEMPT TO PROVE?
The following is a provisional list of things that parallels can prove or at least make an attempt to prove. One aspect of crafting convincing parallels is to make clear what claim is being made on behalf of the parallel and to make sure it is the appropriate claim for the evidence.
1. The meaning of a word or expression (lexicographic parallels).
2. Translation from a particular language (translingual parallels).
3. Direct literary influence (quotation from or use of a specific document).
4. The influence of ideas.
This is a complex and difficult category that can be broken down into a number of subcategories. Provisionally, I suggest:
- i. Familiarity with a certain work or body of literature.
- ii. Familiarity with a certain set of doctrines or a certain viewpoint.
- iii. A shared cultural patterns that helps order the data (e.g., epic themes).
- iv. A shared cultural background. These ideas spread by osmosis throughout a culture and may be found in widely different groups that are members of the culture.
- v. The historical trajectory of an idea or a complex of ideas.
- vi. The life situation and/or the meaning of a difficult text.
5. Membership in a more or less defined social group (such as Essenes, Sadducees, or "sectarians").
6. Membership in a cross-cultural type (e.g., "king," "prophet," or "priest")
7. Participation in a shared phenomenological (cross-cultural) pattern that helps order the data being studied (e.g., someone who shares characteristics with the "prophet" but who also has differences and who functions more generally as an "intermediary").
Much work remains to be done to sharpen our methods of conceiving and applying parallels. This essay is presented in the hope that it advances the discussion.
Bauckham, Richard. "A Quotation from _4QSecond Ezekiel_ in the _Apocalypse of Peter_. _RevQ_ 15 (1992) 437-45.
__________. "Qumran and the Fourth Gospel: Is There a Connection?" In _The Scrolls and the Scriptures: Qumran Fifty Years After_, 267-79. Edited by Stanley E. Porter and Craig A. Evans. JSPSS 26. Sheffield: Sheffield Academic Press, 1997.
Capper, Brian J. "Community of Goods in the Early Jerusalem Church." _ANRW_ II 26.2, 1730-74, esp. pp. 1738-41.
Cross, Frank Moore. "Alphabets and Pots: Reflections on Typological Method in the Dating of Human Artifacts." _Maarav_ 3.2 (1982) 121-36.
__________. "The Epic Traditions of Early Israel: Epic Narrative and the Reconstruction of Early Israelite Insititutions." In _The Poet and the Historian: Essays in Literary and Historical Biblical Criticism_, 13-39. Edited by Richard Elliott Friedman. HSS 26. Chico, Ca.: Scholars Press, 1983.
Davila, James R. "Of Methodology, Monotheism and Metatron: Introductory Reflections on Divine Mediators and the Origins of the Worship of Jesus." In _The Jewish Roots of Christological Monotheism. Papers from the St. Andrews Conference on the Historical Origins of the Worship of Jesus_, 3-18. Edited by Carey C. Newman, James R. Davila, and Gladys S. Lewis. Leiden: Brill, 1999.
__________, "The Dead Sea Scrolls and Merkavah Mysticism." In _The Dead Sea Scrolls in their Historical Context_, 249-64. Edited by Timothy H. Lim, et al. Edinburgh: T&T Clark, 2000.
Esler, Philip F. "Introverted Sectarianism at Qumran and in the Johannine Community." In _The First Christians in their Social Worlds: Social-Scientific Approaches to New Testament Interpretation_, 70-91. London and New York: Routledge, 1994.
Greenspahn, Frederick E. _Hapax Legomena in Biblical Hebrew: A Study of the Phenomenon and Its Treatment Since Antiquity with Special Reference to Verbal Forms_. SBLDS 74. Chico, Ca.: Scholars Press, 1984.
Hendel, Ronald S. _The Epic of the Patriarch: The Jacob Cycle and the Narrative Traditions of Canaan and Israel_. HSM 42. Atlanta, Ga.: Scholars Press, 1987.
Popper, Karl R. _The Poverty of Historicism_. Third edition. New York, N.Y.: Harper and Row, 1964.
Sandmel, Samuel. "Parallelomania." _JBL_ 81 (1962) 1-13.
Smith, Jonathan Z. _Drudgery Divine: On the Comparison of Early Christianities and the Religions of Late Antiquity_. Chicago: University of Chicago Press, 1990.
Speiser, E. A. _Genesis_ (AB; Garden City, N.Y.: Doubleday, 1964)
Suter, David Winston. _Tradition and Composition in the Parables of Enoch_, esp. pp. 14-23. SBLDS 47. Missoula, Mont.: Scholars Press, 1979.
Reproduction beyond fair use only on permission of the author.
Back to the Dead Sea Scrolls Web Page.